Vous êtes sur la page 1sur 44

A Decision-Making Approach
6th Edition

Chapter 1
The Where, Why, and
How of
Data Collection
Fundamentals of Business Statistics - Spring 2006

Chapter Goals
After completing this chapter, you should
be able to:

Know key definitions

What is Statistics
Statistics is the science of data
The Scientific Method
1. Formulate a theory
2. Collect data to test the theory
3. Analyze the results
4. Interpret the results, and make decisions

Example

Exercise: Does the data always conclusively

prove or disprove the theory?

The Scientific Method

The scientific method is an iterative process. In
general, we reject a theory if the data were
unlikely to occur if the theory were in fact
true.

Descriptive statistics

Inferential statistics

Statistical Inference
Statistical Inference
To use sample data to make generalizations
about a larger data set (population)

A Population is the set of all items or

individuals of interest
A Sample is a subset of the population under
study so that inferences can be drawn from it
Statistical inference is the process of drawing
conclusions about the population based on
information from a sample

Testing Theories
Hypotheses Competing theories that we want to test
about a population are called Hypotheses in
statistics. Specifically, we label these competing
theories as Null Hypothesis (H0) and Alternative
Hypothesis (H1 or HA).
H0 : The null hypothesis is the status quo or the
prevailing viewpoint.
HA : The alternative hypothesis is the competing belief.
It is the statement that the researcher is hoping to
prove.

Example
Taking an aspirin every other day for 20 years
can cut your risk of colon cancer nearly in
half, a study suggests. According to the
American Cancer Society, the lifetime risk of
developing colon cancer is 1 in 16.
H0:

H A:

10

You Do It 1.2
(New York Times, 1/21/1997) Winter can give you a cold because it forces
you indoors with coughers, sneezers, and wheezers. Toddlers can give
you a cold because they are the original Germs R Us. But, can going
postal with the boss or fretting about marriage give a person a postnasal drip?
Yes, say a growing number of researchers. A psychology professor at
Carnegie Mellon University, Dr. Sheldon Cohen, said his most recent
studies suggest that stress doubles a persons risk of getting a cold.
The percentage of people exposed to a cold virus who actually get a cold
is 40%. The researcher would like to assess if stress increases this
percentage. So, the population of interest is people who are under
stress. State the appropriate hypothesis for assessing the researchers
theory regarding the population.
H0:
HA:

11

Deciding Which Theory to

Support
Decision making is based on the rare event concept.

Since the null hypothesis is the status quo, we

assume that it is true unless the observed result is
extremely unlikely (rare) under the null hypothesis.
Definition: If the data were indeed unlikely to be
observed under the assumption that H0 is true, and
therefore we reject H0 in favor of HA, then we say
that the data are statistically significant.

12

YDI 1.3
Last month a large supermarket chain received many
customer complaints about the quantity of chips in a
16-ounce bag of a particular brand of potato chips.
Wanting to assure its customers that they were
getting their moneys worth, the chain decided to
test the following hypothesis concerning the true
average weight (in ounces) of a bag of such potato
chips in the next shipment received from the
supplier:
H0:
HA

13

Question

Suppose you concluded HA. Could you be

wrong in your decision? What if you did not
reject H0? Could you be wrong in your
decision?

14

Errors in Decision Making

In our current justice system, the defendant is
presumed innocent until proven guilty. The
null and alternative hypothesis that
represents this is:
H0:
HA :

Truth
H0
HA
based on data

H0
HA
15

Definition
Rejecting the null hypothesis H0 when in fact it
is true is called a Type I error. Accepting the
null hypothesis H0 when in fact it is not true
is called a Type II error.
Note: Rejecting the null hypothesis is usually
considered the more serious error than
accepting it.

16

Type I and II Errors

= Type I error
= The chance of rejecting H0 when in fact
H0 is true
= P(HA|H0)
= Type II error
= The chance of accepting H0 when in fact
HA is true
= P(H0|HA)

17

Whats in the Bag?

Objective To explore the
various aspects of decision
making
Problem statement There are
two identical looking bags,
Bag A and Bag B. Each bag
contains 20 vouchers. The
contents of the bag, i.e., the
face value and the
frequency of voucher
values, are as follows:

Face
Value (\$)

Bag A

Bag B

-1000

10

20

30

40

50

60

1000

Total

20

20

18

Frequency Plot

19

Game Rules

The objective is to pick Bag B.

You will be shown only one of the bags.
You will be allowed to gather some data from the
bag, and based on that information, you must
decide whether to take the shown bag (because you
think that it is Bag B), or the other bag (because you
think that the shown bag is Bag A).
Initially, the data will consist of selecting just one
voucher from the shown bag (without looking into it).
In this case, we say that we are taking a sample of
size n = 1.

20

Example (cont.)
H0 : The shown bag is Bag A
HA : The shown bag is Bag B
Type I error =
Type II error =
Exercise: If the voucher you selected was \$60,
what would you decide? What if the voucher

21

What values of the

voucher (or in what
direction of voucher
values) support the
alternative hypothesis
HA? That is, what is the
direction of extreme?

Face
Chance
Value (\$) if Bag A

Chance
if Bag B

-1000

1/20

10

7/20

1/20

20

6/20

1/20

30

2/20

2/20

40

2/20

2/20

50

1/20

6/20

60

1/20

7/20

1000

1/20

22

Decision Rule 1
Reject the null hypothesis
in favor of the
alternative hypothesis if
the voucher value is
\$50.
Type I error =
Type II error =

23

Summary
Decision Rule Reject H0 if voucher \$50
Rejection Region \$50 or more
We say ... the cutoff is \$50, and larger values
are more extreme

24

YDI: Decision Rule 2

Reject the null hypothesis
in favor of the
alternative hypothesis if
the voucher value is
\$?
Type I error =
Type II error =

25

Why Sample?
A Census is a sample of the entire population

FINISHED FILES ARE THE RESULT OF YEARS OF SCIENTIFIC

STUDY COMBINED WITH THE EXPERIENCE OF MANY YEARS

26

A population or universe is the total elements of interest for a

given problem.
Finite population
Infinite population
A sample is a part of the population under study selected so that
inferences can be drawn from it about the population. Sample
sizes are usually represented by n.
Sampling error (variation) is the difference between the result
obtained from a sample and the result that would be obtained
from a census.
Parameters are numerical descriptive measures of populations /
processes.
Statistics are numerical descriptive measures computed from the
observations in a sample.

27

YDI 2.1
Exercise Nine percent of the US population
has Type B blood. In a sample of 400
individuals from the US population, 12.5%
were found to have Type B blood. Circle your
In this particular situation, the value 9% is a
(parameter, statistic)
In this particular situation, the value 12.5% is
a (parameter, statistic)

28

Good Data?
A sampling method is biased if it produces results that
systematically differ from the truth about the population.
Example Convenience samples and volunteer samples generally
Selection bias is the systematic tendency on the part of the
sampling procedure to exclude or include a certain part of the
population
Nonresponse bias is the distortion that can arise because a large
number of units selected for the sample do not respond.
Response bias is the distortion that arises because of the wording
of a question or the behavior of the interviewer.

29

Example
In the election of 1936 the Literary Digest magazine
predicted that challenger Alf Landon would beat the
incumbent, Franklin Roosevelt. They based their
prediction on a survey of ten million citizens taken
from lists of car and telephone owners, of whom
over 2.3 million responded. This was the largest
response to any poll in history, and based on this,
the Literary Digest predicted that Landon would win
57% to 43%. In reality, Roosevelt won 62% to 38%.
What went wrong? At the same time, a young man
known as George Gallup surveyed 50,000 people
and correctly predicted that Roosevelt would win the
election.

30

YDI 2.3
A study was conducted to estimate the average size of
households in the US. A total of 1000 people were
randomly selected from the population and they
were asked to report the number of people in their
household. The average of these 1000 responses
was found to be 4.6.
1. What is the population of interest?
2. What is the parameter of interest?
3. An average computed in this manner tends to be
larger than the true average size of households in
the US. True or false? Explain.

31

Sampling Techniques
Samples

Probability Samples

Non-Probability
Samples
Judgement

Simple
Random

Convenience

Systematic

Stratified

Cluster
32

Statistical Sampling

Items of the sample are chosen based on

known or calculable probabilities
Probability Samples

Simple

Stratified

Systematic

Cluster

Random

33

Statistical Sampling
A sampling method that gives each unit in the
population a known, non-zero chance of
being selected is called a probability
sampling method (statistical sampling).
Probability Samples

Simple

Stratified

Systematic

Cluster

Random

34

Every individual or item from the population

has an equal chance of being selected

35

Stratified Samples
A stratified random sample is selected by
dividing the population into mutually
exclusive subgroups, and then taking a
simple random sample from each subgroup.
The simple random samples are then
combined to give the full sample.
allows us to obtain information about each
Subgroup
can be more efficient than simple random
sampling

36

Example

37

Systematic Samples
For a 1-in-k systematic sample, you order
the units of the population in some way and
randomly select one of the first k units in the
ordered list. This selected unit is the first unit
to be included in the sample. You continue
through the list selecting every kth unit from
then on.
Convenient
Fast
Could be biased

38

Cluster Samples
In cluster sampling, the units of the population are grouped into
clusters. One or more clusters are then selected at random. If a
cluster is selected, that all units of that cluster are part of the
sample.
Is a cluster sample a simple random sample?
Is a cluster sample a stratified random sample?
Were you to form clusters, how should the variability of the units
within each cluster compare to the variability between the
clusters?
Is this criterion the same as in stratified random sampling?

39

YDI 2.13
Identify the sampling method for each of the following scenarios:
1. A shipment of 1000 3 oz. bottles of cologne has arrived to a
merchant. These bottles were shipped together in 50 boxes with
20 bottles in each box. Of the 50 boxes, 5 boxes were randomly
selected. The average content for these 100 bottles was
obtained.
2. A faculty member wishes to take a sample from the 1600
students in the school. Each student has an ID number. A list of
ID numbers is available. The faculty member selects an ID
number at random from the first 16 ID numbers in the list, and
then every sixteenth number on the list from then on.
3. A faculty member wishes to take a sample from the 1600
students in the school. The faculty member decides to interview
the first 100 students entering her class next Monday morning.

40

Data Types
Data

Qualitative
(Categorical)

Quantitative
(Numerical)

Discrete

Continuous

41

Data Types

42

Key Definitions

A population is the entire collection of

things under consideration

A parameter is a summary measure computed

to describe a characteristic of the population

A sample is a portion of the population

selected for analysis

A statistic is a summary measure computed to

describe a characteristic of the sample

43

Inferential Statistics

Making statements about a population by

examining sample results

Sample statistics
Population parameters
(known)
Inference
(unknown, but can
be estimated from
sample evidence)

Sample

Population